The histogenesis of Ewing's sarcoma remains enigmatic, despite much work to elucidate its origins. We have assumed that Ewing's sarcoma, in its usual state of differentiation, lacks any specific features of known childhood tumors. Certain lines of evidence from other studies, such as the presence of a reciprocal (11:22) chromosomal translocation in Ewing's sarcoma and peripheral neuroepithelioma, and similar patterns of reactivity with panels of monoclonal antibodies, have suggested a possible common histogenesis for these otherwise dissimilar tumors. Since neural tumors in general are known to respond to differentiating agents such as dibutyryl cyclic AMP, nerve growth factor, and retinoic acid by developing features of differentiated neural tissues such as neurites and increased numbers of dense core granules, we have treated a series of Ewing's sarcoma tumor cell lines in vitro with these agents under a variety of conditions, alone and in conjunction with one another. To date, the initial results strongly suggest that at least those tumors which are successfully grown in vitro are intrinsically capable of neural differentiation in response to treatment with these agents. Four of four lines so studied (and reported previously to lack any spontaneous evidence of neural differentiation, even after year of growth in vitro) responded by producing long, slender processes in culture. Ultrastructural examination of these processes revealed dense core granules. Immunocytochemistry with antisera to neuron-specific enolase, an antigen found in neural tissue, was negative prior to treatment but positive afterwards in all four lines. These initial results are being confirmed with other techniques, including catecholamine fluorescence, neurotransmitter enzyme profiles, extracellular matrix synthesis studies, and patterns of monoclonal antibody reactivity.